Merge pull request #11 from baloise/main

Add GUI
baloise · Oct 30, 2024 · 98fe8cc · 98fe8cc
2 parents a2758be + 7b75604
commit 98fe8cc
Show file tree

Hide file tree

Showing 4 changed files with 89 additions and 11 deletions.
diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
@@ -1,4 +1,4 @@
-name: Docker
+name: Docker build & push
 
 # This workflow uses actions that are not certified by GitHub.
 # They are provided by a third-party and are governed by

diff --git a/src/templates/form.html b/src/templates/form.html
@@ -1,14 +1,54 @@
 <!DOCTYPE html>
-<html>
+<html lang="en">
 <head>
-    <title>Simple Form</title>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>YoYo MaskR</title>
 </head>
 <body>
-    <h1>upload text to mask</h1>
-    <form method="post"> 
-        <label for="name">Input:</label><br>
-        <input type="text" id="ipunt" name="input"><br><br>
-        <input type="submit" value="Submit"> 
+    <h3>Input:</h3>
+    <form id="inputForm">
+        <input type="text" id="inputData" placeholder="Enter text to anonymize" required>
+        <input type="submit" value="Submit">
     </form>
+
+    <h3>Response:</h3>
+    <textarea id="responseField" rows="10" cols="50" readonly></textarea>
+
+    <script>
+        document.getElementById('inputForm').addEventListener('submit', function(event) {
+            event.preventDefault(); // Prevent the default form submission
+
+            const inputData = document.getElementById('inputData').value;
+
+            // Use a relative URL for the API endpoint
+            const apiEndpoint = '/api/mask'; // Relative URL
+
+            // Send a POST request to the API
+            fetch(apiEndpoint, {
+                method: 'POST',
+                headers: {
+                    'Content-Type': 'application/json',
+                },
+                body: JSON.stringify({ text: inputData }), // Send the input data as JSON
+            })
+            .then(response => {
+                if (!response.ok) {
+                    throw new Error('Network response was not ok');
+                }
+                return response.json(); // Parse the JSON response
+            })
+            .then(data => {
+                console.log('Success:', data); // Handle the success response
+                // Display the response in the textarea
+                document.getElementById('responseField').value = JSON.stringify(data, null, 2); // Format the JSON response
+            })
+            .catch((error) => {
+                console.error('Error:', error); // Handle any errors
+                // Optionally display the error in the textarea
+                document.getElementById('responseField').value = 'Error: ' + error.message;
+            });
+        });
+    </script>
 </body>
-</html>
+</html>
diff --git a/src/utils/llm.py b/src/utils/llm.py
@@ -8,7 +8,7 @@
 Task instructions: Analyze the text provided after 'Text to anonymize' carefully for all names of persons and places.
 For each name that you find evaluate whether it is a new name or just a repetition or variation of a name you have already found before.
 Names of persons are labeled as #person_1#, #person_2#, etc. Names of places are labeled as #place_1#, #place_2#, etc.
-Only return a json dictionary without any comments or markdown formatting around it.
+Only return a json dictionary without any comments or markdown formatting around it. Do not return keys without values.
 Example input: 'Tony Stark and Peter Parker walk through New York where Peter wants to show Tony the Broadway.'.
 Example output: {{"#person_1#": ["Tony Stark", "Tony"], "#person_2#": ["Peter Parker", "Peter"], "#place_1#": ["New York"], "#place_2#": ["Broadway"]}}
 Text to anonymize: {text}
@@ -31,4 +31,6 @@ def llm_find_entities(text, temperature=0, template=TEMPLATE, raw=False):
     result = chain.invoke({"text": text})
     if raw:
         return result
-    return {k: sorted(v, key=len, reverse=True) for k, v in json.loads(result).items()}
+
+    ret = {k: v for k, v in json.loads(result).items()}
+    return {k: set(v) for k, v in ret.items() if v}
diff --git a/tests/test_llm.py b/tests/test_llm.py
@@ -0,0 +1,36 @@
+import pytest
+from src.utils.llm import llm_find_entities
+
+def test_llm_find_entities_basic():
+    text = "Tony Stark and Peter Parker walk through New York where Peter wants to show Tony the Broadway."
+    expected_output = {'#person_1#': {'Tony Stark', 'Tony'}, '#person_2#': {'Peter Parker', 'Peter'}, '#place_1#': {'New York'}, '#place_2#': {'Broadway'}}
+    result = llm_find_entities(text)
+    assert result == expected_output
+    assert result == expected_output
+
+def test_llm_find_entities_no_entities():
+    text = "This is a text without any names of persons or places."
+    expected_output = {}
+    result = llm_find_entities(text)
+    assert result == expected_output
+
+def test_llm_find_entities_repeated_names():
+    text = "Alice and Bob went to Wonderland. Alice met Bob at the Wonderland park."
+    expected_output = {
+        "#person_1#": ["Alice"],
+        "#person_2#": ["Bob"],
+        "#place_1#": ["Wonderland"],
+        '#place_2#': ['Wonderland park']
+    }
+    result = llm_find_entities(text)
+
+def test_llm_find_entities_raw_output():
+    text = "Tony Stark and Peter Parker walk through New York where Peter wants to show Tony the Broadway."
+    result = llm_find_entities(text, raw=True)
+    print("result")
+    print(result)
+    assert isinstance(result, str)
+    assert "Tony Stark" in result
+    assert "Peter Parker" in result
+    assert "New York" in result
+    assert "Broadway" in result